Topology-Aware Neighborhood Learning for Source-Free Cross-Scene Hyperspectral Image Classification

2026-08-06

AI Summary

This work addresses the challenge of source-data-unavailable cross-scenario hyperspectral image classification by proposing a topology-aware unsupervised domain adaptation framework. The method uniquely integrates global collaborative representation with local nearest-neighbor search to construct a contextual neighborhood topology that comprehensively captures the intrinsic manifold structure of the target domain. To enhance pseudo-label quality, it introduces entropy-momentum-based pseudo-label refinement, complemented by a log-inner-product topological consistency constraint and an information maximization regularizer. Extensive experiments on three standard cross-scenario hyperspectral datasets demonstrate that the proposed approach significantly outperforms current state-of-the-art methods. Ablation studies further validate the effectiveness of each component, underscoring the critical role of topological modeling in source-free hyperspectral domain adaptation.

Abstract

Domain adaptation has advanced cross-scene hyperspectral image classification, significantly improving discriminative capability in complex scenarios. However, privacy rules or storage limits often block access to data from the source domain. Conventional domain adaptation methods become impractical, severely restricting their utility in realistic remote sensing scenarios. To tackle this challenge, we propose a topology-aware source-free learning framework. We first introduce the entropy momentum pseudo-labeling (EMP) to refine k-means assignments by leveraging entropy-aware confidence and temporal prediction momentum. Under the guidance of the refined pseudo-labels, we further utilize the contextual neighborhood topology (CNT) to exploit the intrinsic geometric structure of the target feature space. Combining the global structural information extracted by collaborative representation with the local similarity information modeled by nearest neighbor search, the CNT accomplishes the comprehensive encoding of manifold-level geometric properties in the target domain feature space. The overall objective integrates cross-entropy on refined pseudo-labels, log inner product-based topology consistency, and an information-maximization term for balanced classification, ensuring stable adaptation in the source-free setting. Extensive experiments on three typical cross-scenarios demonstrate that the proposed method exceeds state-of-the-art performance, and ablation studies further validate the contribution of each module. The results highlight the critical role of topology-aware modeling in achieving robust and accurate classification without source data.

Problem

• source-free
• cross-scene
• hyperspectral image classification
• domain adaptation
• topology-aware

Innovation

• source-free domain adaptation
• topology-aware learning
• pseudo-label refinement
• hyperspectral image classification
• manifold geometry